World in 2050

Jonny Axelsson
13 min readSep 5, 2022

What are your best predictions for the world 30 years from now? What will the cities be like? Politics? New infrastructure? Technology? Fashion? Which sectors and regions will grow and which will fall behind?

Bear in mind that even though 2050 is a while into the future, it is not far future. It’s as far ahead of us as 1990 was behind us. Overall 2050 should be pretty much like today, but where will it differ?

The “AI”s are not dramatically more intelligent than they were 30 years ago, but they are certainly far more capable. AI is basically applied statistics, and the big difference is that we have much bigger data sets today, a trend that is going to continue.

And that is a mechanism for a more authoritarian future. Big Data is not now, and is unlikely to be equally shared, and he who controls information controls the universe.

(cue Cercei/Littlefinger dialog here)

Connectivity is the game, and probably will remain so for the next decades and beyond.

In the 1980s connecting two computers (with Ethernet) cost about as much as the computers themselves, putting computer networks in the realm of academic institutions and larger corporations. But after that, the challenge hasn’t been cost, but methods. If you combine a database with another database, you get much more than two databases, but you might spend as much effort doing that as setting up the databases in the first place. So you got protocols, APIs and query languages.

And “AI” works, to a point, in handling these interfaces. But they too are limited and labour-intensive. The “intelligence” is not in those programs, the information is in the data. We can expect significant progress in extracting and integrating data. That, not scale or processing capacity, has been shown to be the hard problem.

This in fact seems to have massive benefits of scale, and benefits of scale tend to map into monopolies, whether governments or huge corporations. That is a valid concern. Or to use the classic XKCD:

Many recent technologies have greatly benefited democracy, several of the upcoming technologies seem to benefit oligarchies.

We have had some impressive improvements in machine motion software, and that is likely to continue. The bottom-up approach is bearing fruit. The Boston Dynamics/Black Mirror horror movies are a little bit smoke and mirrors, but we are clearly going this way and will have that in real products. When machines can manage locomotion, balance, proprioception, spatial orientation, touch and other senses autonomously. It probably won’t affect us much in the 2020s, but in the 30s it will. By then I assume we have also managed to do swarming well, so a group of machines can behave as one swarm. That should be some show.

AI is essentially applied statistics. I “studied” neural networks around 1990 in the sense that I took a graduate course I didn’t finish. It’s a fascinating field, but not something I saw myself doing. A neural network is a multi-layer graph connecting inputs (e.g. photos) with outputs (e.g. names). The graph weighted so input that is most “alike” to the output gets selected. The alikeness is mathematical so we don’t see why they are alike, and they could even back then match as well or better than people, and the perception was different from ours. So they could pick out from noisy signals things we couldn’t see/hear, but also utterly fail two inputs that to our eyes/ears were completely different. Neural network fails are kind of funny.

Changes in last 30 years: Obviously they are getting a lot better. I am not surprised that NN are used for face/category recognition, pattern matching is what they do, but computer vision is very hard (different lighting, angles, picture quality etc), and they have performed much better than could be expected. Maybe most importantly the assumption was that there would be a scaling problem. The NN performed worse when they had to recognise e.g. 500 faces instead of 50. Not so surprising, we too confuse faces (or Chinese characters or what have you) when there are too many of them, so we’d assume you could have a facial recognition system for company security, but not for an entire nation. This is not a performance/processing power issue, the methods must have become much better.

VelesHomais said:

1990 was closer to 1930 than it is to today in terms of day to day life.

Miniaturization and the power of computers along with internet changed our lives more than electricity,cars, trains all combined did.

No, it definitely was not. You forget that some of us were alive in 1990.

The world in 2020 is much closer to the world in 1990 than the world in 1970 was, at least for the West. That’s because 1990 was kind of an inflection point and 1970 was the end of an era, people just didn’t know it yet. 1969 was a busy year. NASA landed on the Moon. Just before that the Internet had uttered its first word. This was the Space Age, that replaced the Atomic Age just a decade earlier. The age of rocket ships. Everything would be bigger and faster, supersonic flights crossed the Atlantic, hovercrafts crossed land and sea. Space stations and Moon bases beckoned. Big Oil and Big Car ruled the roost, and USA ruled the world.

It really didn’t last long. The oil shock came in 1973, the Apple II microcomputer (and Star Wars) came in 1977. The Concorde crash in 2000 was a belated bookend (the plane was financially obsolete by then), like the better timed Titanic crash in 1912 bookended the period of fastest technological progress in history.

In the 1990s we created the world we see in 2020. Instead of “faster, bigger” it was “smaller, smarter”. In this progress was fairly fast until about the financial crisis in 2008. There are signs we are pretty much at the same stage now as we were in 1970.

I specified the Western World. Until 1990 the West, the first world had advanced, while the second and third world had fallen way behind. At 1990 the second world collapsed, and since the third world has caught up. In 1990 Japan was the yellow peril, in 2020 China is, and the rest of Asia is following. As predicted (in the 1990s) Asia is rising. As might Africa eventually (in range of 2050). In 1984 we had “Do they know it’s Christmas?”, today Ethiopia is a regional power.

The weird history of flying cars, 1920–1970 — Rare Historical Photos

Ever since the car was invented, man has dreamt of taking it to the skies. These pictures show the long and not so successful history of flying cars.

goschio said:

Apart from the internet and mobile phones (there were car phones LOL) the whole life was pretty much identical to now. Of course this only applies to developed western countries. I know there was big differences to the life east of the iron curtain.

Hofstadter’s Law: It always takes longer than you expect, even when you take into account Hofstadter’s Law.

The speed of change is not constant (and certainly not accelerating). Mobile telephony as a technology seemingly followed the traditional sigmoid trajectory slow-FAST-slow, but there are two trends in one.

In the 70s (and earlier) mobile telephony was as you said car technology for a niche market. This changed into the 80s yuppie phones.

In this whole period, and through the 90s as well, miniaturization was the game. Until then it was primarily a phone. But it was also in the 90s digitization began, with many forgettable and forgotten false starts.

By the 00s digital features dominates, and by the 10s it is all Google/Apple walled gardens. One development piggybacking on another is common, and like here it is usually intentional, not accidental.

So what will happen to the phone eventually? It will fade into the background, that is the fate of all tech (either obsolete or invisible). From communicator and digital Swiss army knife the “handy” as the Germans call it is morphing into the Me Module. We see that already, whenever you are to identify that you are you or they are yours, largely the phone plays the role.

A decade or two ago it was speculated to be a style icon, a token to display your personality, and for a while it was, and a phone was something to accessorise. But largely the trend has gone in the opposite direction, phones are getting more drab and non-distinct. In an age where we can mass produce in lots of 1, everything can be customised, phones follow the Henry Ford adage “any color you like as long as it is black”. Phones will be back in fashion, just like wrist watch have and hats haven’t.

As all things get pervasive, “smart” and chatty, there is no actual need for phones anymore, but manufacturers will try to manufacture some. So in their life times mobile phones got smaller, then bigger, then disappeared from view.

First of all, there is an attitude change (or paradigm change if you prefer) in what is considered progress. It’s no longer a manned based on Mars (unless you are Bezos, Branson, or Musk, they got a bit of the 1960s mindset), but basically IT. We can kind of predict attitude changes a few decades ahead, because we ourselves are fairly predictable. This list is in good part the world of 2010 as envisioned in 1990, some things didn’t happen, or less. Some happened more or better.

If we sum up our IT story so far:
30s-50s: Theory about computers and radio telephony
50s-70s: Computers, primitive radiophones. Algorithms/software and network theory
70s-90s: PCs and their platforms/programs, early internet and protocols
90s-10s: Web sites and apps, China&Silicon Valley, big data and globalisation

I myself declared that the Internet would be “done” by 2020. What can I say? Hofstadter’s Law. And it is kind of done, I think I did better than those predicting the paperless society or the deaths of cities due to telecommuting. We are pretty paperless by now, but it took an additional 30 years of paper deluge. Likewise, present pandemic excluded, cities are still growing, and will keep doing that for a while.

The Internet can be divided in three rough phases, the virtual internet, the systemic internet, and the physical internet.

The virtual internet was immediate. You may not have discussed with Australians in the 90s, but I did, back in the 80s on Usenet (RIP). Email is fifty years old, newsgroups about forty, and the first web discussion software about twentyfive, but only got good about twenty years ago. Youtube and online gaming is virtual, but capacity was a delay. Banking, stock and the like are technically virtual, but were held back by a number of factors, like the insistence that Java was the only safe protocol/framework.

Not a problem for internet shopping, that actually does involve logistic, but it is easier to pack and ship products than to make banks mend their ways. Amazon’s book delivery system was basically an upgrade of mailorder retail, and Netflx used to ship DVDs in the mail. Logistics have been (partially) reformed, as have business practices, from retail to industry to government (and banking), but there is still some way to go. Changing business practices is a hard and slow process.

Almost 20 years ago I chatted with a lady from the Norwegian tax authorities. Her dream/vision was to be able to deduct VAT and taxes realtime during money transactions. That way you never actually pay taxes as the money has been deducted before you receive them.

For employee taxation it has been like this for a decade, though not for business transactions and not in realtime. In (most? all?) of Europe you get an annual message how much you have paid in taxes and why, and if you think you have a reason to you can contest that. That’s progress, but it doesn’t involve the very areas where most money changes hand and tax evasion is most likely. Perhaps by 2030.

The physical internet is lagging, but that is to be expected. A coffee pot controller was one of the first things that came up on the Web in the 1990s. Everything we make can be “smart”, whether or not there’s any obvious reason to, and they can interact with us and each other. We got internet of things, robotics with actuators and sensors. In real life right now, not that much. The Alexa surveillance system and drone videos are steps on the way, but hardly full metal future. Yet.

The speed of technological advances is uneven, and it can also get lost/misplaced for extended periods of time. It is hard to see how that particular eventuality would happen now, but it is healthy to keep that in mind.

How fast it is also depends on which phase of the process you are looking at: The theory/methodology, the early product developments (usually with at least one false start), or when the tech becomes pervasive. There’s a decade (or more) between the phases. There will be humps with higher activity and developments and troughs where little seems to happen.

To take an example that should move us into the future instead of dwelling on the past: The brain. The 1990s was declared to be the decade of the brain. That was an apt designation. So much happened in the decade that any scientific knowledge before that decade might just as well have been done in the middle ages.

Now we are in the phase we have some niche products based on that research. We got our brain-on-a-chip, we get neural interfaces for spine/nerve damage, early attempts at computer-assisted vision for blind and so on. These products will be far better, but these niches are hardly pervasive. Well before 2050 we should have our convenient cybernetic augmentations.

Which makes for a nice segue from the brain to life expectancy. We expect significant, but not dramatic increases in life expectancy by 2050. Probably something on the order of 5–10 years increase, presumably a global levelling out, lower increases in countries with high expectancy, higher in countries with low expectancy.

Five additional years of life is not to be sneered at, but more importantly the quality of life is likely to continue to improve. I expect we will get most common causes of dementia beat by 2050. Same with many of the other rich world / old age diseases (diabetes et al). Cancer is an arms race, basically we should be able to move up the age at which we’ll die of it, and treating cancer will be less unpleasant, but we will not be able to stop cancer from happening.

This is a fairly cool chart on how life expectancies have increased globally over the years.

Life Expectancy

When and why did the average age at which people die increase and how can we make further progress against early death?

That said, another way to measure life expectancy is the chance of growing up to become 65. And on this metric most of the world has improved greatly, even more so than for life expectancy (radically better health system). But the US is an interesting exception.

:jax: said:

One last extrapolation.

From 2012 to 2017, the percentage of US men growing up to the ripe age of 65 FELL from 80.64% to 79.69%, a fall of 0.19% per year. Meanwhile the world average rose from 71.76% to 73.45%, a rise of 0.34% a year.

So by 2029, if this trend continues, you would have a higher chance of living to the age of 65 if you were born a random place OUTSIDE the US, than if you were born in the US.

And meanwhile, your chances in Sub-Saharan Africa grew from 50.29% to 54.62%, an increase of 0.89%/year. So from the year 2042 on your chances of survival in Sub-Saharan Africa will be better than they are in the US.

All men in the US should keep that in mind…

By that extrapolation by 2050 men in Sub-Saharan Africa will have greater chance of growing up to the age of 65 than men in the US will.

There will be a niche for space tourism, a.k.a. manned space missions. An increasing number of nations see this as a marquee mission, better than hosting Olympic games. Bored billionaires want to make some splash with their cash. So expect a string of missions beyond Earth orbit, a handful in the 2020s, a few handfuls in 2030s, and then either it falls out of fashion by the 2040s (like it did in the 1980s), or lower cost and better life support make it more of a mass tourism thing.

But space belong to machines, not meat bags. Our bodies are not made for space, but machines can be. And they get autonomous enough to do the necessary tasks on their own. But it will certainly be a lossy business in the 20s and 30s, and is unlikely to run break-even by the 40s. That will put a damper on future space exploration, even autonomous.

William Gibson said:

the future is already here — it’s just not very evenly distributed

When Gibson wrote Neuromancer (1984) he had not been on the Internet, but many of us had at the time. He had been to Japan (which many of us hadn’t). Now, China has taken the place in Western psyche that Japan had (and a bigger chunk of the world economy).

When Obama won the US presidential election, China was the land of hope and change, a place simulaneously in the past and in the future. In that it followed the track of a good number of countries, like Japan, and between the two there were four tigers. Thailand didn’t quite make it, and now it is Vietnam’s turn to jump.

Four Asian Tigers — Wikipedia

Then of course, further back, there was Sweden